- Posted by Anthony Harrington, May 19, 2011
- 2 Comments
- Login to add your comment
Anthony Harrington
Quantitative managers who rely on mathematical models to assemble investment portfolios, rather than going out to meet company management and “kicking the tyres”, like traditional “long-only” equity analysts, have faced some very challenging times since the great Wall Street meltdown of 2008. It did not help their cause that much of the crash was blamed on the exotic financial engineering which gave the world CDOs and CDO squareds – instruments that were invented by math types using techniques that are opaque not just to the man and woman in the street, but often to their own senior management as well.
Collateralized Debt Obligations (CDOs)
CDOs, or Collateralized Debt Obligations, have since been dubbed “the banker’s home made bomb” since they were derivative products and the risks associated with the underlying asset class upon which they were based, namely bundles of US sub-prime mortgages, were almost invisible to the buyers of the CDOs. To the buyer, what mattered was the cash flows from the CDOs, and they took heart from the fact that the CDO products were rated AAA by the major ratings agencies.
In the words of Paul Wilmott, a quant who now leads 7city's financial engineering/quant courses, CDOs started out as a brilliant idea. The quants who devised CDOs solved the problem of how to slice up a large parcel of securitized mortgage-backed assets into five or six tranches, with each tranche yielding appropriate levels of reward for appropriate levels of risk. The whole idea of CDOs was that those who got the highest reward got hit first if individual mortgage holders in the securitized bundle defaulted. Only when their equity stake had been reduced to zero would the next highest tranche holder start to suffer, and so on up the chain.
This made the top slices of a CDO, those furthest from the initial defaults, look pretty immune to the vagaries of fate. Statistical techniques, based on huge data sets of mortgage payment histories to provide the underlying normative analysis, were also supposed to make the most risky slice of a CDO a lot more immune to failure.
Subprime crisis
As we now know all too well, and what the quants, for all their mathematical flair, certainly did not know, or did not focus on, was that mortgage underwriting standards in the US sub-prime market had dwindled to virtually nothing. Everyone could put themselves up for a “liar’s loan” and borrow more than 100% of the value of their homes. Entire securitized “parcels” were poisoned with rubbish loans and the quants’ assumptions about risk were wildly off the mark as a result.
Once the defaults started to pour in, it became painfully clear that they were going to be way in excess of anything the quants’ models had told them was even remotely likely. Likelihoods that had been evaluated as a one-in-a-trillion event turned out to be commonplace events. Ouch.
By mid-2007, as the sub-prime fiasco started to unfold, quant proprietary trading desks at major US banks found themselves heading into deeper and deeper trouble. The rest is history. Ex-Wall Street Journal writer Scott Patterson wrote a book that saw to it that quant reputations got pounded in the aftermath. Called “The Quants: How a bunch of backroom geeks took over Wall Street and almost ruined it”, the book told the story of how quant based proprietary trading desks at major investment banks, such as Morgan Stanley’s Process Driven Trading, staffed by “young math whizzes” made their institutions fortunes in the boom years running up to 2007 – and then lost billions through the crash when their models couldn’t find a way out of the carnage.
Patterson’s view of the crash typified the wider investment industry’s sense that quantitative analysis had fallen off its horse in a major way when markets became turbulent. Writing for Bloomberg Markets, Richard Teitelbaum has a great piece on Goldman Sach’s AQR Capital Management, one of the world’s largest hedge funds. He points out that AQR was reported to have seen assets under management fall from a peak of $39.1 billion in September 2007 to $17.2 by March 2009, as clients pulled out of quant funds after the crash.
Clifford Asness, who heads AQR, and is well known for his directness and colourful comments, is reported to have summed up his experience through the crash as: “I heard the Valkyries circling – I saw the grim reaper at my door.”
AQR has since convinced many of its clients that its new models have profited from lessons learned in the crash and are well capable of beating benchmarks and generating consistent returns. Funds under management are on the rise again as AQR funds bear out Asness’s claims and outperform their benchmarks.
Innovation for survival?
The firm’s story, at least as regards clients withdrawing funds in the wake of the crash, has parallels across the quant fund universe. In his piece for the Financial Times, “Quant shops innovate to survive”, Michael Shari cites Barclays Capital analyst Matthew Rothman’s estimate that by December 2009, quant funds had fallen from their peak in June 2007 by 55-60%, to $483 billion in assets under management (AUM). Another $80 billion to $100 billion was withdrawn from quant funds through 2010. In 2011, however, the position has reversed.
Though we are not yet deep enough into the year to quantify the turnaround (no pun intended) quant funds appear to be redeeming themselves in the eyes of institutional investors. Moreover, institutional fund managers who have to make decisions about where to allocate pension fund money have had time to get over their panic and to revisit the quant fund universe to look again at who performed reasonably over the crash and who didn’t. They have also had time to notice that even those funds that didn’t, as we saw with AQR, now seem to have replaced defective models with ones that are more able to deal with non-normal returns and more extreme market swings. We pick this point up in part 2.
Further reading on quant funds, asset allocation and investment:
- Asset Allocation Methodologies, by Tom Coyne
- The Case for SMART Rebalancing, by Arun Muralidhar and Sanjay Muralidhar
- Booms, Busts, and How to Navigate Troubled Waters, by Joachim Klement
Tags: asset allocation , CDOs , investment , investment banks , quantitative investment , quants , statistical arbitrage , Wall Street
Comments
Login or register to post your comments.
www.qfinance.com
Anonymous Comment says:
Wed Jun 20 15:37:28 BST 2012
evoneb says:
Thu May 26 06:14:37 BST 2011